ABSTRACT
Detailed knowledge of cellular networks that are modulated by Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is needed to understand viral replication and host response. So far, transcriptomic analyses of interactions between SARS-CoV-2 and cells were performed on mixed populations of infected and uninfected cells or using single-cell RNA sequencing, both leading to inaccurate or low-resolution gene expression interpretations. Moreover, they generally focused on annotated messenger RNAs (mRNAs), ignoring other transcripts, such as long non-coding RNAs (lncRNAs) and unannotated RNAs. Here, we performed deep polyA+ transcriptome analyses of lung epithelial A549 cells infected with SARS-CoV-2, which were sorted based on the expression of the viral protein spike (S). To increase the sequencing depth and improve the robustness of the analysis, the samples were depleted of viral transcripts. Infection caused a massive reduction in mRNAs and lncRNAs, including transcripts coding for antiviral innate immune proteins, such as interferons (IFNs). This absence of IFN response probably explains the poor transcriptomic response of bystander cells co-cultured with spike positive (S+) ones. NF-{kappa}B and inflammatory response were among the pathways that escaped the global shutoff in S+ cells. In agreement with the RNA-seq analysis, inflammatory cytokines, but not IFNs, were produced and secreted by infected cells. Functional investigations revealed the proviral function of the NF-kB subunit p105/p50 and some of its known target genes, including IL32 and IL8, as well as the lncRNA ADIRF-AS1, which we identified as a novel NF-kB target gene. Thus, analyzing the polyA+ transcriptome of sorted populations of infected lung cells allowed unprecedented identification of cellular functions that are directly affected by infection and the recovery of coding and non-coding genes that contribute to SARS-CoV-2 replication.
ABSTRACT
TWIST1 is a basic helix-loop-helix transcription factor, and one of the master Epithelial-to-Mesenchymal Transition (EMT) regulators. We show that tumor suppressor miR-145-5p controls TWIST1 expression in an immortalized prostate epithelial cell line and in a tumorigenic prostate cancer-derived cell line. Indeed, shRNA-mediated miR-145-5p silencing enhanced TWIST1 expression and induced EMT-associated malignant properties in these cells. However, we discovered that the translational inhibitory effect of miR-145-5p on TWIST1 is lost in 22Rv1, another prostate cancer cell line that intrinsically expresses high levels of the CPEB1 cytoplasmic polyadenylation element binding protein. This translational regulator typically reduces TWIST1 translation efficiency by shortening the TWIST1 mRNA polyA tail. However, our results indicate that the presence of CPEB1 also interferes with the binding of miR-145-5p to the TWIST1 mRNA 3'UTR. Mechanistically, CPEB1 binding to its first cognate site either directly hampers the access to the miR-145-5p response element or redirects the cleavage/polyadenylation machinery to an intermediate polyadenylation site, resulting in the elimination of the miR-145-5p binding site. Taken together, our data support the notion that the tumor suppressive activity of miR-145-5p on TWIST1 translation, consequently on EMT, self-renewal, and migration, depends on the CPEB1 expression status of the cancer cell. A preliminary prospective study using clinical samples suggests that reconsidering the relative status of miR-145-5p/TWIST1 and CPEB1 in the tumors of prostate cancer patients may bear prognostic value.
